Interview approach

Interviews are conducted by researchers in order to gather exhaustive and comprehensive information to explore the experiences, views and beliefs or motivations of individuals on the phenomenon under investigation (Gill et al., 2008; Rowley, 2012). Qualitative methods, such as interviews are considered as the most appropriate tool for exploring sensitive topics (e.g. ineffectiveness of the drug delivery process) where detailed participants‘ insights are required or little is known about specific matters (Holloway & Wheeler, 2013). A representative example is the study conducted by Rossetti et al., (2011); they adopted an interview approach to collect data in order to identify the major forces that impact upon the biopharmaceutical supply chain. Similarly, Bhakoo and Choi (2013) used a semi-structured interview protocol to gather data related to healthcare personnel‘s reaction to institutional and endogenous pressures for technology implementation. Kay and Blinkhorn (1996) used interviews as research methods in order to collect information about the factors that influenced GPs' decision on treatment choices.

The fundamental purpose of the research interview is to ask questions related to the study phenomenon that are likely to yield as much information as possible and address the research aims and objectives (Roulston, 2010). This can be achieved by choosing the appropriate interview form; there are three fundamental diverse forms of research interview classified based on their level of ‗structure‘: unstructured, semi-structured and structured (Rowley, 2012).

Unstructured interviews are used when researchers attempt to encourage participants to talk around a theme; during the natural flow of communication, questions are spontaneously generated. This type of interview might start with an opening question such as ‗what is your role within the PSC?‘ and then the discussion will develop from the initial response. Their use is generally considered where little is known about the subject area and it might generate more structured questions (Bryman, 2001). Unstructured interviews are usually time-consuming and their management and guidance require skill and experience (Roulston, 2010).

118 Conversely, structured interviews are considered as verbally administered questionnaires that include a list of well-structured questions (Silverman, 2010). This list of questions is used with every interviewee with little or no variation. The answers expected are usually short and as a result, structured interviews are often less time-consuming than unstructured interviews. In addition, this type of interview is easier to administer and facilitate.

In the middle of the spectrum stands the semi-structured form of interviews. This is the most common form of interview as it can be designed in a more flexible manner, enabling the researcher to drive the discussion in order to gather the required data (Mason, 2002). Although this interview format consists of several key questions aiming to define and gain a better understanding of the study context, it also allows the interviewees to focus on and analyse a specific subject, generating new views and topics (Rowley, 2012). Excessive interactivities are developed between the interviewees and the researcher, which allows the interviewer to extract more information (Kvale & Brinkmann, 2008).

In this thesis, unstructured and semi-structured research interviews were conducted in order to understand the drug delivery process in both Greece and the UK through the participants‘ experiences, opinions, views and values. The interview approach was adopted because it is considered the most appropriate research tool for data gathering where there is insufficient knowledge about the study subject and the potential participants might be more familiar with this approach. The following sub-sections discuss in detail the interviews design and the sample approach which are core factors in the success of a data collection approach.

Interviews’ design

In this research, the qualitative data were collected by conducting unstructured and semi-structured research interviews with key professionals working within the downstream domain of the PSC in Greece and the UK. The initial interviews (N=8) were unstructured, including more general themes derived from the literature review, focusing on understanding the delivery process and the pharmacies‘ role in it. Subsequently, those interviews were analysed, which enabled the researcher to further develop the following semi-structured interviews; this set of initial interviews had an exploratory character informing the following diverse data collection procedures: semi-structured interviews and survey questionnaire. Therefore, the following semi-structured interviews (N=22) involved well-structured themes that helped the researcher to drive the discussion and gather the

119 required information. In-depth interviews provide comprehensive and exhaustive data, which can generate new directions in social sciences (Denscombe, 2007; Reige, 2003).

The research interviews included a list of open-ended questions which enabled interviewees to discuss and analyse the research topic based on their knowledge, experience and beliefs (Silverman, 2010). The list of questions was divided into three parts: i) the first part referred to general questions about the study phenomenon and the role of the interviewee therein; ii) the second part included specific questions regarding themes (e.g. the factors preventing an effective delivery process) identified by reviewing reports and previous research; iii) finally, the third part focused on the interviewees‘ personal views and beliefs on whether the drug delivery process could be improved through innovation. Each of the interviews was informed by the existing literature and the analysis of the previous interviews, which enabled the data to evolve over time in a direction that addressed the research aims and objectives.

Before conducting the semi-structured interviews, the researcher created a list of potential participants and restructured the questions, included in the interviews based upon their particular expertise. Pilot testing of the interview questions was essential to ensure that the questions were clearly stated and understandable (Rowley, 2012). The interview questions used in this thesis were developed and the pilot tested by five academics related to the study subject. Their main suggestions were related to the terminology used; for example they pointed out that, terms such as: Lean, supply chain or reverse logistics might not be familiar to the participants. Therefore, the questions were changed accordingly to minimise the risk of misunderstanding. Subsequently, the researcher initially contacted the potential participants through emails or phone calls to introduce herself and the research project and check their willingness and availability. Ethical principles such as anonymity and confidentiality were also explained, as these might have increased the likelihood of participation and openness of the interviewees.

Sample population

Sampling techniques are used by researchers in order to collect the required data focusing only on a specific group of cases (Saunders et al., 2009). Researchers select the sample that matches a number of criteria and best answers the research questions, meeting the research aims and objectives (Matthews & Ross, 2010). In this thesis, the target sample involves professionals working in hospital and community pharmacies in two diverse geographical areas: Greece and the UK. This

120 research aims to explore how innovative programmes could improve the downstream domain of the PSC which is directly related to patients, identifying the related issues. Therefore, only those specialists who work within this domain could be considered as potential participants.

In particular, in 2013, approximately 410,000 pharmacists were operating in the EU; the number of professionally active pharmacists reported in the majority of the EU member states was 50-106 per 100,000 inhabitants (Eurostat, 2015b). Greece recorded a high number of professionally active pharmacists, at 106 per 100,000 inhabitants in 2013 (approximately 11,600 pharmacists) (Eurostat, 2015b); the vast majority of them, about 80%, were working in independent-community pharmacies (Vozikis et al., 2015). On the other hand, this number in the UK, in 2013, was 80 per 100,000 inhabitants (approximately 51,600 pharmacists) (Eurostat, 2015b), with approximately 70% of those operating as independent-community pharmacists (NHS England, 2013). From these pharmacists only those working within the particular study area, the downstream domain of the PSC, could be considered as potential participants in the current study. Unfortunately, the exact number of the potential participants is not available.

As previously mentioned, during the qualitative data collection, the researcher faced difficulties in approaching the potential interviewees. They were very cautious about being involved in this research, mainly because they thought that their role was not related with Operations Management (OM) and Supply Chain Management (SCM) practices. They were therefore excluded due to concerns related to the minimum knowledge about these particular practices and organisational performance (e.g. Nulty, 2008). In addition to this, some of the potential participants were reluctant to take part in this study because of their heavy work load. The researcher approached only those specialists that could be reached in terms of geographical distance. Although the described issues illustrate the magnitude of the challenge faced by the researcher, finally, 8 unstructured and 22 semi-structured interviews were undertaken. Particularly, 5 unstructured and 16 semi-structured interviews were conducted in the UK, and 3 unstructured and 6 semi-structured interviews took place in Greece. Those interviews provided enough data to generalise the qualitative research outputs, as the last interview did not add any consequential data. This ensured that the main research content was covered and, thus the saturation level was reached (O‘Reilly & Parker, 2012; Walker, 2012)

A snowball sampling technique or a network referral sampling was conducted to approach the target sample; it is an efficient technique for accessing hard-to-reach segments of the population (Atkinson & Flint, 2001). This technique enabled the researcher to approach potential interviewees

121 through colleagues or friends, a fact that, on one hand, might have increased the trust between the researcher and the interviewees and on the other hand might have affected the participants‘ opinion of the research subject due to the exchange of knowledge. However, the existence of bias in the results is limited because every single interview was developed differently based on the interviewees‘ expertise and experience. In addition, at the end of each interview, the researcher asked the interviewees to recommend some of their contacts who could potentially agree to be interviewed as well; this was another dimension of the snowball sampling approach that was equally important for increasing the access to data.

After creating a list of potential participants, the researcher contacted them through email or phone calls to provide a brief about the current research and arrange a meeting with them. Aiming to motivate respondents, a report of the future research outputs was offered. The 30 unstructured and semi-structured interviews undertaken varied in their length; the minimum length of interviews was 30 minutes and the maximum length was 90 minutes. Table 3.1 provides an overview of the conducted unstructured interviews and Table 3.2 presents a summary of the conducted semi- structured interviews. Each interview was audio-taped and transcribed verbatim before the data analysis took place.

# Date Reference Position Interview Type

1 25/10/2013 1/UK Lead Pharmacist Unstructured

2 08/11/2013 2/UK Chief Pharmacist Unstructured

3 21/11/2013 3/UK LPC Secretary Unstructured

4 12/03/2014 4/UK Procurement and Homecare

Manager Unstructured

5 12/03/2014 5/UK Chief Pharmacy Technician Unstructured

6 12/02/2015 15/Gr Hospital Pharmacist Unstructured

7 16/02/2015 17/Gr Community Pharmacist Unstructured

8 17/02/2015 18/Gr Hospital Pharmacist Unstructured

122

# Date Reference Position Interview Type

1 15/03/2014 6/UK Acting Chief Pharmacist Semi-structured 2 24/03/2014 7/UK Chief Pharmacy Technician Semi-structured

3 02/04/2014 8/UK Lead Pharmacist Semi-structured

4 04/04/2014 9/UK Community Pharmacist Semi-structured

5 04/07/2014 10/UK Chief Pharmacist Semi-structured

6 09/09/2014 4/UK Procurement and Homecare

Manager Semi-structured

7 09/09/2014 5/UK Chief Pharmacy Technician Semi-structured

8 04/11/2014 11/UK Lead Pharmacist Semi-structured

9 03/02/2015 12/UK Reader Advancing Clinical Practice Semi-structured 10 09/02/2015 13/UK Senior Lecturer, Nursing & Health

Studies Semi-structured

11 12/02/2015 14/Gr Hospital Pharmacist Semi-structured 12 12/02/2015 15/Gr Hospital Pharmacist Semi-structured 13 13/02/2015 16/Gr Community Pharmacist Semi-structured 14 20/02/2015 18/Gr Hospital Pharmacist Semi-structured 15 21/02/2015 17/Gr Community Pharmacist Semi-structured 16 21/02/2015 19/Gr Community Pharmacist Semi-structured 17 09/03/2015 20/UK Community Pharmacist Semi-structured 18 18/03/2015 21/UK Community Pharmacist Semi-structured 19 23/03/2015 22/UK Community Pharmacist Semi-structured

20 17/04/2015 1/UK Lead Pharmacist Semi-structured

21 05/05/2015 2/UK Chief Pharmacist Semi-structured

22 19/05/2015 23/UK Community Pharmacist Semi-structured

Table 3.2: The overview of the conducted semi-structured interviews

Data analysis

Cresswell (2007) characterised qualitative data analysis as a spiral because the researcher might need to go through the data more than once before they reach the research output. There is no universal recipe for analysing the data; the method adopted is dependent upon the data collected and

123 the research aims and objectives (Saunders et al., 2015). Rowley (2012, p. 268) stated that ―there

are a number of key components of data analysis, including: organising the data set; getting acquainted with the data; classifying, coding, and interpreting the data; and, presenting and writing up the data‖. In order for this analysis process to be achieved, researchers focus on the

meaning of collecting data, trying to identify the key themes.

Thematic analysis has been widely used as a foundational method for analysing qualitative data (Guest et al., 2012). Braun and Clarke (2006, p. 82) highlighted that ―thematic analysis provides a

flexible and useful research tool, which can potentially provide a rich and detailed, yet complex account of data‖. This tool enables researchers to identify, report and analyse themes within the

collected data. Boyatzis (1998, p.63) defined themes as ―the most basic segment, or element, of the

raw data or information that can be assessed in a meaningful way regarding the phenomenon‖.

Themes organise a group of repeating ideas which allows researchers to answer the research questions (Vaismoradi et al., 2016).

There are numerous reported articles referring to thematic analysis agreeing that there is no right or wrong way to conduct it (Tuckett, 2005; Salda a, 2013; Vaismoradi et al., 2016). In this research, the analysis of the research interviews was conducted using Braun and Clarke‘s (2006; 2013) linear model for carrying out thematic analysis. Their model includes six procedures: i) Familiarisation with the data; ii) Generation of initial codes; iii) Searching for themes; iv) Reviewing themes; v) Defining and naming themes; and vi) Producing the final report. In particular, having transcribed the recordings, a list of thematic codes was generated; this process was influenced by the conceptual framework, deductively informed by the literature review and the author‘s research interest. Basit (2003, p. 144) highlighted the importance of the coding process by stating that ―codes or categories

are tags or labels for allocating units of meaning to the descriptive or inferential information compiled during a study‖.

Researchers can develop a thematic analysis by either using computer software, such as NVivo or analysing the qualitative data manually using Word documents (Rowley, 2012). Although the use of computer software packages could facilitate the qualitative data analysis (Basit, 2003), in this thesis, the thematic analysis was undertaken using MS Excel. The use of computer software saves time, avoiding the tedious and frustrating process of manual analysis (Winsome & Johnson, 2000). However, the risk associated with these packages relates to their effects on research. Winsome and Johnson (2000, p.395) listed the concerns of using these packages: ―a focus on quantity instead of

124

retrieval methods, distancing of the researcher from the data, inappropriate use of technology, time consumed in learning to use computer packages, pressures or expectations that all qualitative researchers will use them, and increased commercialism‖. Computer software packages have been

developed under a certain epistemology (Coffey et al., 1996), which does not necessarily fit with the purposes of the study (Petty et al., 2012). In light of this and avoiding losing control of the data, the thematic analysis has been conducted manually. The developed themes related to specific issues observed during the delivery process are derived from the participants‘ experiences and opinions. These themes will be presented and analysed in the following Analysis Chapter (Chapter Four). The following section discusses the second phase of the research design which includes the analysis of the quantitative data. The reason for using a survey questionnaire is explained and in addition to this the questionnaire design, the sample characteristics and the quantitative data analysis are described.